I was told that in romaji the long vowel of o would be ō (see this tutorial)

However, I am confused by these two words. In Google Translate they both were romanized as Tōri (see here), but their 平仮名 are different

  • 通り とおり
  • 東リ とうり
  • 2
    I guess that's a limitation of romaji. Note that there are several sytems for writing romaji. It's not standardised. You could write these as 'toori' and 'touri' (I have no idea what 東リ is) and there would be no ambiguity. Best to move on from romaji as soon as you can and forget about it. It woudl also help to clarify your question. Not sure what you actually want to know. Feb 16 at 10:47
  • Thanks! I guess I sometimes use Google Translate as a dictionary, and then use their romaji as a guide to Japanese IME (under Windows). I guess in cases like this I should also consult a real dictionary. I did not know about the romaji limitation before. Thanks for your insight.
    – leeyuiwah
    Feb 16 at 10:54
  • 1
    @user3856370 The difference between とう and とお isn't a romaji issue though, but rather a matter of which kana transliteration is considered correct. You can try and "forget about" romaji, but that won't do you any good when you're staring at a keyboard, and wondering if you need to type t-o-o-r-i or t-o-u-r-i for the IME to show you 通り (the latter is considered incorrect and hence won't work).
    – Will
    Feb 16 at 12:27
  • 1
    @EddieKal Dude, I'm typing this on a Japanese keyboard. My point was that regardless whether you type t-o-o-r-i or と-お-り on your keyboard, you need to know that the second vowel is a お/o for the desired kanji prediction of 通り to show up.
    – Will
    Feb 16 at 15:44
  • 1
    @Will, and for the original poster -- if you're interested in typing with diacritics like Ø, ç, or ā, and you're on Windows, there's a freeware app called WinCompose that I've used for years to good effect. Full disclosure: I'm only a user who has reported a couple issues to the developer. The project page on GitHub: github.com/samhocevar/wincompose Feb 17 at 0:17

Responding to the extended discussion in the comments --


The second vowel is //o// in both 通【とお】り and 東【とう】. (東【とう】り is not a word, so I won't repeat that here.) When you're talking about vowels, you're talking about pronunciation. The second kana is either お or う, but in both 通【とお】り or 東【とう】, the vowel is //o//.

Kana spellings

This gets really involved. Hop to the bottom if the details don't matter to you. 😄

The kana spelling difference is due to different origins -- historically, 通り came from older とほり. Medial (mid-word) //h//'s often lenited (softened) and disappeared in many words, leaving behind the vowels. The pronunciation shifted before the kana spelling did -- the とほり kana spelling was the "correct" spelling up through the spelling reforms of the mid-20th century.

This older historical kana spelling is still included in many monolingual Japanese dictionaries, since modern readers are still likely to encounter this. For instance, see the 通り entry at Kotobank. At the top of the page, you'll see kana spellings listed like とおり〔とほり〕. The kana string 〔in brackets〕 is the historical spelling.

Historical spellings for native Japanese terms -- kun'yomi

If you spend some time with a monolingual Japanese dictionary that includes these, you'll notice that many entries for kun'yomi terms with vowel-vowel combinations like //oo// or //au// come from older historical spellings with an //h// or //f// in between.

All of the modern //h// or //f// kana, はひふへほ, were probably originally pronounced with an initial //p//, as //pa pi pu pe po//. These lenited over time to become first the bilabial fricative //ɸ//, kinda like an "f". We know that these "f" pronunciations at the beginnings of words were still current in 1603, when the Japanese-to-Portuguese dictionary the Nippo Jisho was published. We can see this in the scanned copy over at Google Books, in entries like on this page for F ANTES DO I ("F before I"), with entries like:

Fi. Sol. (Fi. Sun.)

Meanwhile, in a medial (mid-word) position, these pronunciations lenited further to often become a //w// sound. We see that in places like the word 川, modern かわ and historical かは. The Nippo Jisho entry here (right-hand column, sixth from the top) shows:

Caua. Rio, ou ribeira. (Kawa. River, or stream.)

In pretty much all cases except before //a//, this //w// sound basically disappeared -- so Old Japanese //opo// ("big; many", the root of modern 大【おお】きい or the prefix 大【おお】) became early Middle Japanese //ofo//, then maybe mid-Middle Japanese //owo//, and then recorded late Middle Japanese //oː//.

Historical spellings for Chinese borrowings -- on'yomi

Entries for on'yomi terms generally don't have this, and instead will show kana indicating different vowel values than we have in the modern pronunciation -- mostly things like あう where the modern language has おう. Historically, these were previously pronounced with an //au// diphthong (two-vowel sound), which was probably pronounced a bit like English ow!. This gradually flattened from //au// to a long-vowel //ɔː// sound a bit like English awww, and then this //ɔː// shifted further to become modern //oː//, the long-O sound in everyday Japanese.

These historical kana spellings -- and indicated pronunciations -- for on'yomi terms point towards the phonological shapes of the words when first borrowed from Middle Chinese. 東 was borrowed originally with a pronunciation probably like //tou// with that //u// on the end, and this flattened to modern //oː//. See the Chinese pronunciation section over at Wiktionary. While the reconstructed Middle Chinese reading is shown as //tuŋ//, I've noticed after much research that the modern Min Nan readings are sometimes closer to the Japanese readings than the Middle Chinese reconstruction -- possibly due to Min Nan being closer to the original dialect of Middle Chinese from which the readings were borrowed into Japanese. One of the Min Nan pronunciations for 東 is listed as tong, and final //ŋ// in Chinese often correlates to an う in the Japanese on'yomi.

Meanwhile, 刀 was borrowed as //tau//, and this shifted to modern //toː//. See the 刀 entry in Daijisen at Weblio, where again we see two kana spellings for the on'yomi, listed as トウ(タウ). The (タウ) in parentheses is the historical kana spelling. And the Middle Chinese pronunciation for this was reconstructed as //tɑu//.

Conclusion about guessing kana spellings from pronunciations

Without knowing the origins of the term, you just can't really tell whether a long-O sound //oː// uses お or う as the second kana.

If you know at least whether the pronunciation is on'yomi or kun'yomi, it's often a good bet that long-O kun'yomi -- in the middle of any word, and not just the volitional / "let's" verb ending -- is spelled with an お for the second kana, while long-O on'yomi is spelled with an う for the second kana.

  • In case you've missed it, 東リ is a real company name, and its official English name is TOLI. (リ is in katakana and it originally stood for Linoleum.)
    – naruto
    Feb 17 at 1:25
  • @naruto, excellent detail. I maintain that it's still not really a "word" per se, but it's good to know that it's not just at typo. :) Feb 17 at 9:27

Different writing systems are useful for different things.

In English, if we wanted to, we could write things down the way we pronounce them. The best way to do that would be using a special alphabet called the International Phonetic Alphabet (IPA). For example, if I wanted to write down how I say knight using IPA, I would write /naɪt/. We don't usually write English that way, but we could.

Instead, we usually write things down roughly as they would have been pronounced in Middle English. Our pronunciation has changed a lot, but our spelling has only changed a little, so the two drifted apart. We write knight, but we don't pronounce that ‹k› or ‹gh› anymore. Our spelling shows us where the word knight came from, and how it used to be pronounced.

Something similar happened in Japanese. Over the centuries, the spoken language changed, but the system of kana spelling didn't change much, and the two drifted apart. This changed after World War II, when Japan adopted spelling reforms which brought kana spelling very close to actual pronunciation. Certain aspects of the old spelling system were kept intentionally, however!

For example, let's look at 王 'king' and 追う 'chase':

  • 王 'king' is pronounced [オー]{HL}, but is written おう in kana. There is no う in the pronunciation, but it's written with う anyway. This reflects the origin of the word, which comes from Chinese. When a word or morpheme from Chinese has a long /oː/ sound, the second half is spelled う.
  • 追う 'chase' is pronounced [オウ]{LH}. It is written おう in kana, and the う makes sense because it's actually pronounced that way.

Again, different writing systems are useful for different things. When I write 王 or 追う, people who can read Japanese can tell those words apart instantly. If I wrote おう, you'd only be able to tell them apart from context. When I write [オー]{HL} or [オウ]{LH}, I'm transcribing the spoken language instead of writing the words the usual way.

So what if we want to write Japanese with Latin letters ("romanization")? What should we do? The answer depends on what your goals are:

  • If we want to copy the kana spelling using Latin letters, we should write ou for both words. Let's call this transliteration.
  • If we want to write down the spoken language and focus on how it's pronounced, we could write ō for 王 and ou for 追う. Let's call this transcription.

Different people at different times have had different goals, and so different systems of romanization have shown up over time.

In your examples, the spelling tōri is used for both words. Why? Because it's a transcription of the spoken language, reflecting how the words are pronounced; it's not intended to be a transliteration, reproducing the kana spelling.

And where does the kana spelling come from? Well, in 東リ, the 東 part comes from Chinese. It's just like 王, where the second half of the long /oː/ sound is spelled in kana with a う. And that's because the post-war spelling reforms stopped short of actually respelling these words with an お, despite their pronunciation. In other words, the う is a holdover from the past.

But 通り isn't from Chinese. It's a native Japanese word, and it never had a う sound historically, so there was never a reason to write it with that kana. Just like the う spelling, the お spelling reflects where the word came from.

So really, it's all just a matter of different writing systems being useful for different things.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.